4 4.2 Key Observations Human Rights Implicated: Right to freedom of expression, opinion and information; Right to privacy            All examples on Burmese social media were written in Burmese, with one exception. 88% of examples contained language directed primarily at the Muslim community. 12% of these fitted the criteria of the most dangerous forms of speech: • There were several examples of a powerful or influential speaker who capitalised on a fear of the audience, including calls to action such as violence against a community where there have been previous episodes of inter-communal violence. • These examples were shared widely on social media, potentially reaching millions of users. • However, while this could be considered an influential means of dissemination, social media is not the sole or primary source of news in Myanmar. All of the samples suggested the audience faced danger from Muslims, either a threat that Muslims are becoming too dominant in society, or that they are spoiling the integrity of Myanmar, for example by marrying Buddhist women. 38% of samples included a call to action, either to discriminate (e.g. by avoiding Muslim-owned shops and businesses), hostility (e.g. that Muslims should be denied citizenship or ethnic minorities should be driven out of the country) and actual calls to kill Muslims. The researcher considered 30% of samples to have fake profile names and were therefore anonymous or not identifiable. For example, several user names were recognisable as names of characters in Burmese novels, or translated into English as phrases like “a beautiful evening”. Some user names had more intimidating translations, for example “the person who guards ethnicity” or phrases intended to be insulting to Muslims. Around 10% of samples compared Muslims to certain animals. The Dangerous Speech Framework includes, as part of the variable on ‘speech’, referring to people as other than human (e.g. vermin, pests, insects or other animals) as an attempt to de-humanise the victim and one indicator of violence. 17% of samples used language or symbols specific to Myanmar, such as images of someone or something being stepped on, considered an insult in Burmese Buddhist society, or using the style of Buddhist teaching or proverb in a derogatory way to Muslims. The posts that were shared most widely were quotes by well-known Burmese figures, links to news articles or alleged accounts of killings of Buddhists by Muslims (all unconfirmed), or calls to boycott Muslim-owned shops and businesses. The posts that received the most reaction/response were those made by a politician or religious leader. One politician alleged a Muslim had set a school on fire, which was shared 1,300 times. The same politician advocated the burning of a mosque if it was built in a particular area, and received over 1,000 positive responses. A religious leader’s post encouraging people not to give housing to Muslims received 1,300 positive responses and was shared 830 times. When influential figures, such as a politician or religious leader, made statements against Muslims, supporting comments by normal users were the most violent of the samples, including calls to kill Muslims. • In the recorded examples, explicit calls to kill Muslims were posted as comments 146 PAGE CHAPTER 4.2: HATE SPEECH